ducpy.search.image_ocr
======================

.. py:module:: ducpy.search.image_ocr

.. autoapi-nested-parse::

   Server-side image OCR helpers for DUC search.

   Uses ``rapidocr-onnxruntime`` with custom PP-OCRv6_tiny ONNX models when the ``ocr`` extra is installed, otherwise
   OCR is gracefully skipped.



Attributes
----------

.. autoapisummary::

   ducpy.search.image_ocr.logger
   ducpy.search.image_ocr._RAPID_OCR_ENGINE
   ducpy.search.image_ocr._server_side_ocr_available
   ducpy.search.image_ocr._extract_image_text_with_ocr


Functions
---------

.. autoapisummary::

   ducpy.search.image_ocr.server_side_ocr_available
   ducpy.search.image_ocr._get_local_cache_dir
   ducpy.search.image_ocr._get_rapid_engine
   ducpy.search.image_ocr.extract_image_text_with_ocr


Module Contents
---------------

.. py:data:: logger

.. py:data:: _RAPID_OCR_ENGINE
   :type:  Any | None
   :value: None


.. py:function:: server_side_ocr_available() -> bool

   Return whether the ``ocr`` extra (RapidOCR) is installed and usable.


.. py:function:: _get_local_cache_dir(repo_id: str) -> str | None

.. py:function:: _get_rapid_engine() -> Any

.. py:function:: extract_image_text_with_ocr(image_bytes: bytes, *, ocr_language: str) -> tuple[str, bool]

   OCR an image using the ``ocr`` extra, or skip when unavailable.


.. py:data:: _server_side_ocr_available

.. py:data:: _extract_image_text_with_ocr